View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Trepo - Institutional Repository of Tampere University This is an Accepted Manuscript of an article published by Portland Press. Biochemical Society Transactions. 2017.45(2):475-464. Available online: http:// dx.doi.org/10.1042/BST20160272. Targeting ROR1 identifies new treatment strategies in hematological cancers Hanna Karvonen1, Wilhelmiina Niininen1, Astrid Murumägi2 and Daniela Ungureanu1 1 BioMediTech, BMT, University of Tampere, 33014, Tampere, Finland 2 Institute for Molecular Medicine Finland, FIMM, University of Helsinki, 00290, Helsinki, Finland Corresponding author name and email: Daniela Ungureanu email: [email protected] Abbreviations: ALL, acute lymphocytic leukemia; B-ALL, B-cell acute lymphoblastic leukemia; BCR, B- cell receptor; BTK, Bruton tyrosine kinase; CLL, chronic lymphocytic leukemia; mAb, monoclonal antibody; MCL, mantle cell lymphoma; NHL, Non-Hodgkin lymphoma; ROR1/2, Receptor tyrosine kinase-like orphan receptor 1/2; Wnt, wingless-related integration site Abstract ROR1 is a member of the ROR receptor family consisting of two closely related type I transmembrane proteins, ROR1 and ROR2. Due to mutations in their canonical motifs required for proper kinase activity, RORs are classified as pseudokinases lacking detectable catalytic activity. ROR1 stands out for its selective and high expression in numerous blood and solid malignancies compared to a minimal expression in healthy adult tissues, suggesting high potential for this molecule as drug target for cancer therapy. Current understanding attributes a survival role for ROR1 in cancer cells, however its oncogenic function is cancer-type specific and involves various signaling pathways. High interest in ROR1 targeted therapies resulted in the development of ROR1 monoclonal antibodies such as cirmtuzumab, currently in phase I clinical trial for CLL. Despite these advances in translational studies, the molecular mechanism employed by ROR1 in different cancers is not yet fully understood, therefore more insights into the oncogenic role of ROR1 signaling are crucial in order to optimize the use of targeted drugs. Recent studies provided evidence that targeting ROR1 simultaneously with inhibition of BCR signaling is more effective in killing ROR1-positive leukemia cells, suggesting a synergistic correlation between co-targeting ROR1 and BCR pathways. Although this synergy has been previously reported for B-ALL, the molecular mechanism appears rather different. These results provide more insights into ROR1-BCR combinatorial treatment strategies in hematological malignancies, which could benefit in tailoring more effective targeted therapies in other ROR1-positive cancers. ROR family of pseudokinases ROR family of proteins belongs to the non-canonical Wnt signaling pathway and comprises of two type I, single-pass transmembrane glycoproteins, ROR1 and ROR2, closely related to MuSK (muscle-specific kinase) and Trk (tropomyosin) family receptors [1, 2]. The extracellular region (ectodomain) contains an immunoglobulin-like domain (IgG), a cysteine-rich domain (CRD) with sequence and structural similarities to the CRD of Frizzled proteins followed by a Kringle domain. A single transmembrane helix connects the ectodomain to the cytoplasmic region, which contains a tyrosine kinase-like or pseudokinase domain followed by two serine/threonine (Ser/Thr) rich domains flanking a proline (Pro) rich domain (Figure 1A). ROR ligands were initially unknown, hence their “orphan” denomination, however subsequent studies have demonstrated that both ROR1 and to a greater extent ROR2 are able to bind Wnt5a presumably via their extracellular CRD domain [3–7]. ROR Kringle domain, a highly-folded cysteine-rich domain that mediates protein-ligand interaction in coagulation proteins, appears to play a role in ROR1/ROR2 heterodimerization and subsequent activation of Rho/Rac1 GTPase pathway in leukemia cells [7]. The cytoplasmic tyrosine kinase-like domain lacks several conserved residues important for phosphotransferase activity, therefore RORs are classified as pseudokinases (Figure 1B). Although early studies reported evidence of autocatalytic activity for both ROR1 and ROR2 [1, 8, 9], the prevailing view of RORs is as inactive pseudokinases, since using purified recombinant proteins the nucleotide binding and autocatalytic activity could not be detected for either of ROR [10, 11]. X-ray crystallographic studies of unphosphorylated ROR2 pseudokinase domain revealed an activation loop that occludes the substrate- and ATP- binding sites via the unique position of Tyr555 side chain, a feature not observed in other kinases [12]. Sequence alignment of ROR1 and ROR2 pseudokinase domains shows high-level of homology (Figure 1B), indicative of similar functional mechanisms. Interestingly, restoration of glycine-rich loop to its canonical consensus sequence does not result in an increased in vitro catalytic activity for purified ROR cytoplasmic domains, suggesting a predominantly non-catalytic function for these pseudokinases [10]. Functional evidence for ROR1 catalytic activity came from studies using mutation of the invariant Lys506 in 3-sheet to create a kinase dead enzyme. ROR1 K506A mutant could not sustain downstream signaling and growth advantage in lung adenocarcinoma cells suggesting that ROR1 kinase activity is required for cell survival [13]. Similar studies have shown that ROR1 could trans-phosphorylate HER3 at a previously unidentified Tyr1307 site and this phosphorylation was abolished by using ROR1 K506A kinase dead mutant [14]. While K506A mutation may alter ROR1 structural integrity and compromise its downstream signaling without affecting its so- called kinase activity, the possibility of a very low or atypical enzymatic activity could exist. The insolubility of both RORs K506R mutants as recombinant proteins (cytoplasmic domain) compared to fully soluble wild-type proteins leave us unable to provide a clear biochemical evidence for the effect of this mutation on ROR catalytic properties in vitro [10]. Phosphorylation of ROR1 has been observed in cell lysates by using either ROR1 phospho-specific or pan-phosphotyrosine antibodies, and several kinases including Src and Met were shown to directly interact with and phosphorylate ROR1, demonstrating that ROR1 takes part in the dynamic network of intracellular phosphorylation events either as a substrate or presumably as an active kinase [13, 15–17]. While ROR1/ROR2 heterodimerization occurs in response to Wnt5a binding, ligand-induced ROR1/ROR2 trans- phosphorylation or auto-phosphorylation have not been detected thus far [7, 10]. The ability of ROR1 to phosphorylate exogenous substrates needs to be investigated beyond K506A mutant overexpression, ruling out the possibility that intermolecular interactions could be responsible for the observed functional effect. The Pro-rich domain of ROR1 serves as substrate for Met phosphorylation in cells with Met- amplification and as a binding site for Src kinase, while recent reports have added new functions for this domain as binding site for 14-3-3ζ and HS1 (hematopoietic cell-specific Lyn substrate 1) to mediate downstream signaling leading to cell proliferation and leukemia cell migration [13, 15, 16, 18, 19]. On the other hand, ROR2 Pro-rich domain serves as a binding site and substrate for CKI kinase resulting in serine-threonine phosphorylation of the second Ser/Thr domain followed by subsequent tyrosine phosphorylation of ROR2 pseudokinase domain [3]. The functional role of the C-terminal Ser/Thr domains in ROR1 awaits further studies. ROR family of proteins are evolutionally conserved and share relatively high level of homology from Drosophila to mammals, underlining their important biological functions [20]. In line with this, mROR1−/− mice show postnatal growth retardation and have a reduced life expectancy, whereas mROR2−/−mice die at birth due to severe skeletal and heart abnormalities [21–23]. mROR1/mROR2- deficient mouse showed a more severe phenotype than mROR2−/− alone, indicating that mROR1 and mROR2 are functionally redundant during embryonic development [24] . The strong ROR1 expression during normal embryonic and fetal development is not sustained in all postpartum healthy tissues. Low ROR1 levels were detected in adipose cells and to a lesser degree in the pancreas, lung and a subset of intermediate B cells [25–27]. Moreover, using a more sensitive ROR1-specific monoclonal antibody recent studies have shown relatively high expression levels of ROR1 in parathyroid, pancreatic islets and multiple regions in the gut [28]. However, upregulation of ROR1 is commonly detected in both hematological and solid tumors [29], which attracted high scientific interest to explore the functional advantage of targeting ROR1 expression in cancer cells as a therapeutic strategy. ROR1 and hematological cancers ROR1 overexpression was initially identified on B-cells from chronic lymphocytic leukemia (CLL) but not on normal B-cells from healthy donors, suggesting that ROR1 expression could become a marker for CLL [25]. Furthermore, several research groups have identified ROR1 but not ROR2 expression in many other hematological malignancies, indicating that ROR1 is more prevalent in blood cancers. ROR1 upregulation was detected in primary samples of mantle cell lymphoma (MCL), marginal zone lymphoma (MZL), diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL), and high